Extraction of cellular networks and microenvironmental characteristics in liver cancer tissues using geospatial approaches for therapeutic and diagnostic potential

Liver cancer arising in the context of metabolic dysfunction-associated steatotic liver disease (MASLD) has been increasing in prevalence in recent years. These tumors are characterized by remarkable heterogeneity in both malignant cells and the surrounding stromal and immune compartments, which contributes to therapeutic resistance and disease progression. Understanding the spatial organization and cellular interactions within the tumor microenvironment (TME) is therefore essential for developing more effective therapeutic strategies. In this study, we performed single-cell RNA sequencing (scRNA-seq) and Visium spatial transcriptomics analyses using human liver cancer specimens derived from MASLD-associated cases to dissect intra-tumoral heterogeneity at both the molecular and spatial levels. We developed a spatial analytical framework that integrates pathway activity scores with geostatistical approaches and distance metrics derived from histopathological landmarks, calculated using a digital unroll method. This approach enabled the identification of region-specific cellular populations and spatial gradients of metabolic and immune activities within tumor tissues. Furthermore, we employed a cell-cell communication analysis to delineate localized signaling networks and microenvironmental niches that sustain tumor progression. Distinct interaction patterns were observed among hepatocyte-like cancer cells, endothelial cells, and immune infiltrates, suggesting spatially restricted communication hubs. Notably, we identified specific secreted factors from defined cell types that may serve as potential non-invasive biomarkers reflecting intratumoral spatial states. Collectively, our study provides an integrative geospatial framework to map the cellular architecture and communication networks in MASLD-associated liver cancer. This spatially resolved understanding of tumor ecosystems may contribute to precision stratification of patients and the development of novel therapeutic strategies.